17:00:41 <nickm> #startmeeting weekly network team meeting
17:00:41 <MeetBot> Meeting started Mon Feb 13 17:00:41 2017 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:41 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:00:42 <nickm> hi all
17:00:47 <nickm> it's monday again!
17:00:51 <dgoulet> yellow!
17:00:54 <ahf> hi
17:00:56 <nickm> we'll start with status reports as usual
17:01:04 <isabela> o/
17:01:07 <nickm> Last week I tried to get the stable branches full of backports,
17:01:13 <nickm> tried to get consensus on #21278 fixes
17:01:17 <nickm> helped with cor dev hiring
17:01:29 <nickm> tried to get consensus on #20894 fixes
17:01:35 <nickm> brainstormed with ahf about sponsor4 stuff
17:02:06 <nickm> tried to get more people to migrate away from 0.2.9.<=8 .  The total share of the network by weight went down from 12.45 to 10.5%,  so meh.
17:02:27 <nickm> This week I'm hoping I can finally get those two tickets mentioned above merged and fixed and etc
17:02:27 <isis> good morning
17:02:32 <nickm> and wrap up hiring
17:02:44 <nickm> and open 031 maybe
17:02:55 <nickm> and put out some stable releases
17:03:01 <nickm> and do more sponsor4 design and planning.
17:03:23 <nickm> Reminder!  Next week I'll be with my inlaws, and intermittently off-net
17:03:42 <nickm> Topics for later discussion: Gsoc, 031 triage status, 030 postmortem
17:03:46 <nickm> end for me
17:03:48 <nickm> hi isis
17:03:51 <nickm> who'd like to go next?
17:04:34 <ahf> i can do that
17:04:42 <ahf> Hello, last week I worked on our measurement code for sponsor4, got the first small patches reviwed (but aren't landed yet), continued with measuring overhead (cell, TLS frames), studied a lot of the internals around our connection code (and asked a ton of questions about that), went over the spec process and re-read tor-spec and some of dir-spec, and have now started writing down my thoughts on directory
17:04:48 <ahf> compression negotation schemes. This week will
17:04:50 <ahf> continue with the above + I have the triage duty, so let me know if I do something incorrect there. Over.
17:05:11 <ahf> + a lot of non-tech stuff related to contracts with Erin and some overall tor stories with Nick :-)
17:05:17 <ahf> that's it
17:06:16 <nickm> ok, who's next?
17:06:22 <nickm> dgoulet / asn / isis / isabela ?
17:06:30 <isabela> i can go
17:06:41 <isabela> done with sponsor 4 monthly reports
17:06:46 <isabela> working on final report for sponsors
17:06:54 <isabela> and invited ppl to join me and hiro this friday on irc
17:07:10 <isabela> to talk about features for her trac-cli utility called lightsaber
17:07:12 <isabela> o/
17:07:39 <isabela> i sent this invite to tor-project email list
17:07:49 <isabela> but this was something the network team discussed a while back
17:08:00 <dgoulet> woot
17:08:16 <isabela> done!
17:08:28 * dgoulet is ready to go
17:08:54 <nickm> go
17:08:59 <dgoulet> Status report for me. Last week was a crazy amount of prop224 work on #20657.  It's getting to a point of a "yeah this branch is big". Currently at 27 commits and in the 4k+ new lines of code.
17:09:04 <dgoulet> It has good test coverage so far and I believe it's good progress. It's really the biggest piece of prop224 so that amount of work was expected.
17:09:07 <dgoulet> That being said, you'll have to let me know if you need 030 stuff from me because I believe I have no ticket assigned to me for that milestone now. And that counts for other things you would like from me :).
17:09:09 <dgoulet> I do plan to continue this week the same thing. Also big thanks to asn here that is helping me with pieces of this work.
17:09:11 <dgoulet> --
17:10:25 * isis can go
17:10:29 <nickm> woo!
17:10:50 <nickm> (go)
17:11:47 <isis> last week i continued to work on the deliverables for OTF, mostly the spec describing the rbridge modifications, decided that there are problems getting decaf working for curve25519 and it's too much of a rabithole for now
17:12:16 <isis> this week i will continue working on that and hopefully also set up domain fronting for bridgedb
17:12:19 <isis> done
17:13:21 <nickm> asn?  anyone else?
17:14:20 * meejah has a quick note?
17:14:23 <nickm> yo!
17:14:24 <mikeperry> I am still working on prop#254 in mikeperry/adaptive-padding-draft
17:15:43 <mikeperry> is prop#264 going into 0.3.0 or 0.3.1? I missed that
17:15:52 <nickm> it already went into 0.2.9
17:15:52 <mikeperry> (or later)
17:15:59 <mikeperry> ah, great
17:16:28 <mikeperry> ok, I will be using it for 254 then
17:16:35 <nickm> cool
17:17:03 <nickm> meejah: ?
17:17:28 <meejah> i've been doing some exit scanning stuff
17:17:28 <nickm> dgoulet: oh, I just checked, and all current 030 issues are basically assigned to me
17:17:39 <dgoulet> hehe ok
17:18:15 <meejah> if anyone wants to review the code, it's approaching a shape that's reviewable. I don't want to release it publically though, so DM me if you'd like to review
17:19:30 <nickm> sure, neat
17:19:30 <dgoulet> cool
17:19:33 <meejah> (it's about 600 lines of Python; Twisted + txtorcon etc)
17:20:06 <nickm> if no more updates, moving on to discussion?
17:20:34 <nickm> so, for 030, it's mostly on me right now to try to fix stuff, and on others to try to review stuff
17:20:50 <nickm> so that's a good sign
17:21:20 <nickm> I'm a little unhappy that it won't be into TorBrowser alphas for a couple of weeks, so there might be pervasive client issues we don't know about
17:21:59 <nickm> for 031 please remind me to create branches in a couple of days.
17:22:03 <nickm> once I have all my 030 patches done
17:22:32 <nickm> For gsoc -- did everybody who can mentor have  achance to list themselves as possible to mentor on that link I sent around?
17:22:39 <isis> what are the large things we're trying to land in 031?
17:22:43 <nickm> atagar would like a list of what projects we're suggesting asap
17:23:01 <nickm> isis: mike's previous dataflow branch, prop#224, initial sponsor4 directory improvements, ...
17:23:22 <dgoulet> https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CoreTorReleases#Goalsfornextreleaseseries0.3.1
17:23:31 <dgoulet> we should try to settle or/and expand this ^
17:23:41 <nickm> isis: yes, that list that dgoulet just posted.  But that's not final; it's all suggestions
17:23:57 <isis> awesome, thanks!
17:24:05 <ahf> Wide CREATE cells is for some of the PQ handshake proposals?
17:24:27 <nickm> ahf: for all of them, really.  None of them fits into a CREATE cell unless we add support for wide CREATEs
17:24:41 <ahf> err, yes, shouldn't have said 'some' there.
17:24:43 <nickm> And we kind of need support for wide EXTEND cells before we can do wide CREATE.
17:24:45 <ahf> cool!
17:24:51 <nickm> or rather, split extend cells
17:25:07 <nickm> there's a proposal -- prop#249
17:25:11 <nickm> I don't know if it's right though
17:25:37 <nickm> what, if anything else, should we do to track our large 031 goals?
17:25:40 <nickm> isabela: any thoughts?
17:25:47 <nickm> this could be a good guideline for us to triage tickets on.
17:25:55 <isis> #249 seemed correct when i went through it to work on #270
17:26:02 <nickm> Also, we should think about which proposals if any we need to write , revise, or evaluate
17:26:14 <isabela> nickm: 031 will be 3 months too?
17:26:53 <mikeperry> I kind of think that we should ensure all our variable length stuff is actually a multiple of the cell size when sent on the wire
17:26:57 <dgoulet> https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CoreTorReleases#Future
17:27:06 <dgoulet> isabela: 6 months, 3 months merge window, 3 months stabilize
17:27:09 <mikeperry> otherwise, circuit fingerprnting will become trivial
17:27:25 <isabela> dgoulet: so it will end around the time y'all has to be done with sponsorR stuff?
17:27:42 <nickm> isabela: that's the current plan, think?
17:27:45 <dgoulet> isabela: yes :), prop224 has to be over by then to make our sponsor happy :)
17:27:51 <nickm> develop today through april, freeze in May, stable in Aug?
17:27:54 <dgoulet> (at least that,s the goal)
17:28:20 <ahf> is this the normal flow that have been used for the last couple of releases? the 3:3 split?
17:28:28 <nickm> we're still learning IMO
17:28:29 <ahf> so right now we are the end of a stabilization period?
17:28:30 <ahf> ok
17:28:33 <isis> mikeperry: you mean for prop#249? we assume the handshakes are distinguishable from each other and (of course) from NTor
17:28:37 <isabela> nickm: design work for sponsor4 wont be done in time for the code of it be done by freeze right?
17:28:40 <nickm> but that's a good question for the 030 postmortem discussion :)
17:28:50 <dgoulet> ahf: well kind of at the start of the stabilizie 030 :)
17:28:57 <nickm> isabela: I think that some easier parts will be mergeable in 031, and some will take longer.
17:29:09 <dgoulet> ahf: Apr 24, 2017 is still kind of far :)
17:29:09 <isabela> ok
17:29:26 <ahf> dgoulet: ahhh, cool, so it overlaps. thanks
17:29:27 <isabela> those are the types of questions that i think we should use to guide us on triaging for 031
17:30:29 <mikeperry> isis: prop#249 is not fully clear what happens to the last 62 byte EXTEND2 cell in section 5 (is it filled to CELL_SIZE, or not?)
17:31:08 <nickm> isabela: ok.  It would also be good to know if we have other sponsored work that we need to get done on a September-ish timeframe, for the 031 planning
17:31:31 <isabela> i wonder if mikeperry and isis has things they want in for their sponsor work
17:32:49 <mikeperry> isabela: my sponsor2 stuff needs to be studied, ideally on a live tor network, but it doesn't *have* to be in 031 to do that
17:33:03 <isabela> mikeperry: cool
17:33:21 <dgoulet> mikeperry: hey we do have a testnet :)
17:33:28 <isis> mikeperry: the padding from the CREATE2 cell (in §2) is dropped upon converting to EXTEND2? I think? you're right it's not clear
17:34:15 <nickm> isis, mikeperry: A patch for that would be very welcome
17:34:24 <isis> mikeperry: are you worried that some adversary could tell the difference between handshakes?
17:34:37 <nickm> mikeperry: isis : If you can edit https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CoreTorReleases to mention your sponsors and work that target 0.3.1, that would be great
17:34:59 * nickm just edited it to mention a possible 0.3.2 schedule, and to tweak some 031 dates
17:35:02 <isis> nickm: ack, thanks!
17:35:05 <mikeperry> dgoulet: oh woah, like a live one? where are the docs?
17:35:16 <dgoulet> mikeperry: yes like a real Tor 2 :)
17:35:18 <jvsg> isis: I'll be trying to implement prop#270 during this year's GSoC (if I get selected). Is there anything that could be a roadblock for it's implementation?
17:35:28 <dgoulet> mikeperry: docs is not public because we don,t want the node public ;)
17:35:31 <isis> there are no public docs because we don't want people running onions on it
17:35:34 <dgoulet> mikeperry: but I can help you out on this
17:36:28 <nickm> so ... gsoc?
17:36:40 <isis> jvsg: neat, hi!
17:37:04 <nickm> https://storm.torproject.org/shared/MsK3U_cb12cNPqyC5uPvqRYLYwsls_LFDr0s9Q3CdV6
17:37:22 <isis> jvsg: there's a patent roadblock on the crypto, we'll need to rewrite that proposal to use NewHope Simple
17:37:32 <nickm> If anybody would sign up to be my co-mentor on the as-yet-un-co-mentored things, that would be grand
17:37:39 <nickm> I can't IRL mentor more than 1.5 projects.
17:38:03 <nickm> also it would rock to have more projects, more ideas,  more stuff to do, more people to help.
17:38:30 <mikeperry> dgoulet: (thanks, that will be nice. I will let you know when I'm ready. still need to write more unit tests and implement the actual negotiation)
17:40:11 <nickm> I'm going to assume that by the end of today, we have all the gsoc ideas we're going to get, because atagar has really been patient with us
17:40:31 <nickm> 2nd-to-last topic
17:40:33 <ahf> when is it gsoc is running from and to?
17:40:47 <ahf> ~late june to medio august'ish?
17:40:53 <nickm> ahf: https://developers.google.com/open-source/gsoc/timeline
17:40:58 <ahf> thanks
17:41:10 <nickm> (sorry for running through topics so fast --t rying to finish before next meeting starts :)
17:41:29 <nickm> so, wrt the 030 postmortem....
17:42:25 <nickm> should we try to get that all figured out in 20 minutes, or just make some progress today?
17:43:00 <dgoulet> nickm: what do we have to figure out?
17:43:18 <nickm> I think, "what if anything do we do differently in 031 and beyond?"
17:43:31 <nickm> and also "what do we have to make absolutely sure we _don't_ change?"
17:44:03 <dgoulet> (no asn here kind of sad but we can try in 15 min :)
17:45:00 <nickm> huh.  Maybe let's try for next week then.
17:45:05 <nickm> oh wait I'm not here next week
17:45:08 <nickm> let's try via email? :)
17:45:15 <dgoulet> yeah we can do that by email
17:45:21 <nickm> ok
17:45:26 <nickm> sooo -- anything else for this week?
17:45:35 * dgoulet is good
17:45:45 <nickm> (Oh!  And also, thanks for all the Tor, everybody!)
17:45:50 <isabela> just dont forget about friday discussion on features for trac-cli
17:45:52 <isabela> o/
17:46:02 <nickm> #endmeeting